# BIS Bank Lending and Demographics

*Tests whether cross-border bank lending (BIS LBS) represents a missing channel
in the net/gross paper's IIP-based analysis.*

## Test 1: Demographics → Bank Lending Positions

| Variable | Z₁ | SE | p | Z₂ | p | N | Countries | R² |
|:--|--:|--:|--:|--:|--:|--:|--:|--:|
| Outward bank claims/GDP | 3.5013* | 1.9292 | 0.070 | -0.1705 | 0.126 | 839 | 27 | 0.023 |
| Inward bank claims/GDP | 0.6691 | 0.4515 | 0.138 | -0.0370 | 0.174 | 3148 | 73 | 0.016 |
| Net bank claims/GDP | -0.3906 | 0.3967 | 0.325 | 0.0356 | 0.139 | 3150 | 73 | -0.007 |

## Test 2: OECD vs Non-OECD

| Subsample | Variable | Z₁ | p | N | R² |
|:--|:--|--:|--:|--:|--:|
| OECD | out | 4.6286** | 0.043 | 755 | 0.022 |
| OECD | in | 2.7015** | 0.012 | 1580 | 0.013 |
| OECD | net | -0.0149 | 0.987 | 1582 | -0.016 |
| Non-OECD | out | -0.8529 | 0.496 | 84 | 0.462 |
| Non-OECD | in | 0.3643 | 0.229 | 1568 | 0.073 |
| Non-OECD | net | -0.5355 | 0.149 | 1568 | 0.008 |

## Test 3: Z₁×KAOPEN on Bank Lending

| Variable | Z₁ | p | Z₁×KAOPEN | SE | p | N |
|:--|--:|--:|--:|--:|--:|--:|
| out | 1.1912 | 0.107 | 0.0110 | 0.0334 | 0.742 | 728 |
| in | 0.5032** | 0.028 | 0.0160** | 0.0069 | 0.020 | 2923 |
| net | -0.6952*** | 0.006 | -0.0087 | 0.0078 | 0.264 | 2923 |

## Test 4: Does Bank Lending Control Change Core Findings?

*Same-sample comparison (toolkit item 5.8) to avoid composition artifacts.*

| DV | Baseline Z₁ | p | Same-sample Z₁ | p | + Bank ctrl Z₁ | p | Attenuation | N |
|:--|--:|--:|--:|--:|--:|--:|--:|--:|
| ca_gdp | 12.19*** | 0.000 | 18.51*** | 0.000 | 19.11*** | 0.000 | +3% | 3115 |
| income_balance_gdp | -15.45*** | 0.001 | -17.86*** | 0.000 | -17.42*** | 0.000 | -2% | 2933 |

## Synthesis

*Assessment of whether bank lending data changes net/gross paper findings.*

---
*Generated: March 8, 2026. Data: BIS LBS (1977-2025), net_gross_panel.csv.*